Non-psychotic relatives of schizophrenic probands often manifest attenuated psychopathology of this illness including negative symptoms that are hypothesized to be due to decreased brain dopamine (DA) function. We now have systematic evidence suggesting that plasma homovanillic acid (pHVA), an indicator of brain DA activity, is reduced in such relatives in relation to the characteristic negative symptoms that indicate a genetic diathesis to schizophrenia. These pHVA findings are not due to peripheral factors suggesting that they may reflect brain DA activity. These data suggest that the reduced DA activity (assessed by pHVA measures), presumably reflecting a genetic diathesis to schizophrenia, is detectable in non-psychotic relatives. This putative low-DA trait can be used as a biological phenotype in genetic studies to ultimately locate and identify gene(s) underlying the DA dysfunction in schizophrenia. However, before considering such work, the low-DA phenotype needs to be better established. The proposed work is designed to address this current crucial limitation. When using pHVA measures as a phenotype in genetic studies, the well-known limitation of pHVA (inability to localize abnormality in the brain) is no longer an issue, because brain localization is not required in a genetic approach. This instrument is highly stable and suitable for large samples needed for genetic studies. In pilot data, it identified DA-affected relatives with high specificity and sensitivity. In the proposed work, two pHVA measures (i.e., decreased pHVA and decreased morning pHVA decline) suggested by the previous data would be further tested in 90 non-psychotic relatives (from multiplex families with schizophrenia) and 90 normal comparison subjects (without a family history of schizophrenia) in a four-year study. On study days, pHVA will be measured and negative symptoms blindly assessed. Whether the two proposed pHVA measures distinguish relatives from normal subjects under optimized conditions (cross-sectionally and longitudinally) will be determined. The relationship of negative symptoms with pHVA will be further established. Using these two pHVA measures, a method to classify relatives as DA-affected or unaffected will be developed with high sensitivity and specificity. Stability of pHVA measures and of the classification will also be established. Once the low-DA phenotype has been established, the door will be open to many important genetic, high risk, and biological possibilities. For example, it would set the stage to employ genetic studies to ultimately locate and identify the gene(s) responsible for the DA dysfunction in schizophrenia. Also, identification of relatives with the DA phenotype at an early age may be very useful in developing preventive strategies even before DA-related gene(s) are identified. This is particularly relevant since many successful interventions affecting DA function are already available.